Premium
Reconsidering the S pecified T etrad Test
Author(s) -
Garcia Karen,
Ennis John M.,
Prinyawiwatkul Witoon
Publication year - 2013
Publication title -
journal of sensory studies
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.61
H-Index - 53
eISSN - 1745-459X
pISSN - 0887-8250
DOI - 10.1111/joss.12060
Subject(s) - assertion , test (biology) , statement (logic) , population , statistics , computer science , noise (video) , tetrad , mathematics , psychology , artificial intelligence , epistemology , philosophy , demography , ecology , biology , programming language , mathematical physics , sociology , image (mathematics)
Additional noise from additional stimuli has been shown to result in a loss of operational power in sensory difference tests. Because the S pecified T etrad test requires the evaluation of four stimuli and has only a slight theoretical power advantage over the two‐alternative forced choice (2‐ AFC ), E nnis and J esionka have argued that the S pecified T etrad test should not be used. But this theoretical assertion had not been confirmed experimentally. In this article, we refute this argument using results from a large‐scale comparison of the 2‐ AFC with the S pecified T etrad test. Using T hurstonian analysis, we quantified the sensory effect size as measured by both tests and found that the S pecified T etrad test is significantly more sensitive than the 2‐ AFC in this setting. In fact, if the sample values from our experiment could be taken as population values, the results of this experiment predict that the S pecified T etrad test is operationally more powerful than a double‐replicated 2‐ AFC . While further investigation of the S pecified T etrad test is needed before such a strong statement can be confirmed, these results indicate that the S pecified T etrad test is worthy of such investigation. Practical Applications Among the many tools in the sensory scientist's tool box are the specified difference testing methods. These methods can be used when the attribute that defines the difference between the samples is known. Historically, the 2‐ AFC has been considered the operationally most powerful method of specified testing, but the 2‐ AFC has never been compared with the S pecified T etrad test. In this article, we provide such a comparison, and show that the S pecified T etrad test deserves attention as a specified testing method.